When a build fails or completes, drone starts killing any containers that are still running. The resulting logs can be confusing to read (especially when trying to debug why a build is failing) because they contain symptoms of containers being stopped, such as dropped network connections. It would be really nice if the logs contained a note which pointed out what point in time drone started killing containers - that way when debugging one could often ignore any logs that came after that point.
Thanks for the feedback. Would it be possible to provide us with a simple / stripped down yaml that demonstrates the deficiency? I think it would help us better design a solution if we can experience this first hand. Thanks!